Comparative risk assessment is an evaluation process designed to rank
environmental problems based on the severity of potential hazards. The
purpose of this paper is to provide an effective statistical approach
to analyze perceived environmental risks. Environmental problems, eva
luative criteria, and other potential moderator variables need to be d
etermined first, and then the risk perception data collected. Repeated
measures analysis is used to first test for interactions between envi
ronmental problems and potential moderator variables. If there are no
significant interactions, then the risk difference among environmental
problems is tested unconditionally; otherwise the risk difference is
tested conditionally. Cluster analysis for environmental problems is p
erformed only when the risk difference is significant. The clustering
results can be objectively determined by using the simultaneous T-conf
idence intervals. Risk-based priority setting is made according to the
clusters obtained. To illustrate this approach, an empirical study of
comparative socioeconomic risks in Taiwan was conducted. Socioeconomi
c impacts areas including social security, quality of life, production
cost, investment willingness, and economic resources are used as eval
uative criteria. Results indicate that selected impact areas do affect
relative risk differences among 24 environmental problems, and the di
fference is significant for each area. Therefore, cluster analysis is
conducted separately for each impact area. Risk-based priority setting
s for clusters of environmental problems are reported.